Dynamic optimal distribution method and system based on power network service resources

Abstract

The invention discloses a dynamic optimization distribution method and system based on power network service resources, and relates to the technical field of power grid resource management, and the method comprises the steps: carrying out the energy flow unification of multiple service resources of a power network, and abstracting the multiple service resources into a resource space; uploading the target service task to a resource management platform for semantic solution coding, and determining an allocation mode set; driving a resource allocation plug-in, identifying an allocation mode set, performing slice reconstruction on a resource space by using an information flow under each allocation mode, executing a multi-thread allocation decision under mode parallelism, determining a resource allocation strategy, and performing service resource management on a power network. The technical problems that existing resource allocation is based on a static allocation strategy of a single resource type, is difficult to adapt to dynamic requirements of multiple service resources, and is low in resource utilization rate and slow in response are solved, efficient management of the multiple service resources is realized through traffic unification and dynamic optimization allocation, and the resource utilization rate is improved. And the resource utilization rate and the response speed are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of power grid resource management, and particularly relates to a dynamic optimization allocation method and system based on power network service resources. BACKGROUND

[0002] With the development of smart grids and energy internet, power networks not only contain traditional power resources, but also integrate various service resources such as computing power, storage and communication. These resources have significant differences in physical properties and functions, which leads to great challenges in resource management. In the prior art, resource allocation is mostly based on single type of resource, and lacks a unified quantitative framework, which is difficult to adapt to the dynamic demand of multi-service resources. In addition, the traditional method mostly adopts a static allocation strategy, which cannot quickly respond to changes in task demand, has low resource utilization and poor system stability. SUMMARY

[0003] The application provides a dynamic optimization allocation method and system based on power network service resources, which is used to solve the technical problems that the existing resource allocation is mostly based on static allocation strategy of single resource type, is difficult to adapt to the dynamic demand of multi-service resources, has low resource utilization and slow response.

[0004] In a first aspect, the application provides a dynamic optimization allocation method based on power network service resources, which comprises: performing energy flow unification on multi-service resources of a power network and abstracting the multi-service resources into a resource space, wherein the multi-service resources at least include power, computing power, storage and communication; uploading a target service task to a resource management platform, determining a set of allocation modes by performing semantic calculation coding, wherein each allocation mode is composed of information flow-value flow; driving a resource allocation plug-in embedded in the platform, identifying the set of allocation modes, slicing and reconstructing the resource space under each allocation mode, taking the value flow as a decision-making guide, executing multi-thread allocation decision-making under parallel modes, fitting allocation results, determining a resource allocation strategy, and performing service resource management on the power network.

[0005] In a second aspect of the present application, a dynamic optimization allocation system based on power network service resources is provided, and the system comprises: an energy flow unification module, which is configured to unify multiple service resources of a power network into an energy flow, and abstract the multiple service resources into a resource space, wherein the multiple service resources at least include power, computing power, storage, and communication; an allocation mode determination module, which is configured to upload a target service task to a resource management platform, determine an allocation mode set by performing semantic calculation coding, wherein each allocation mode is composed of an information flow and a value flow; and a multi-thread allocation decision module, which is configured to drive a resource allocation plug-in embedded in the platform, identify the allocation mode set, slice and reconstruct the resource space according to the information flow in each allocation mode, perform multi-thread allocation decision under parallel execution mode, fit an allocation result, determine a resource allocation strategy, and manage service resources of the power network.

[0006] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0007] The dynamic optimization allocation method and system based on power network service resources provided in the present application relate to the technical field of power grid resource management, and abstract multiple service resources in a power network into a resource space through energy flow unification, determine an allocation mode set by using semantic calculation coding, drive a resource allocation plug-in to slice and reconstruct the resource space according to an information flow, perform multi-thread allocation decision, realize dynamic optimization management of power network service resources, and solve the technical problems that existing resource allocation is mostly based on static allocation strategies of a single resource type, it is difficult to adapt to dynamic demands of multiple service resources, resource utilization is low, and response is slow. The technical effects of realizing efficient management of multiple service resources and improving resource utilization and response speed are achieved through energy flow unification and dynamic optimization allocation. BRIEF DESCRIPTION OF DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0009] Figure 1 A dynamic optimization allocation method based on power network service resources provided in the embodiments of the present application is shown in the flowchart.

[0010] Figure 2 A dynamic optimization allocation system based on power network service resources provided in the embodiments of the present application is shown in the structural schematic diagram.

[0011] Explanation of reference signs: energy flow unification module 11, allocation mode determination module 12, multi-thread allocation decision module 13. DETAILED DESCRIPTION

[0012] The application provides a dynamic optimization allocation method and system based on power network service resources, which is used to solve the technical problems that the existing resource allocation is mostly based on static allocation strategies of single resource type, it is difficult to adapt to dynamic demand of multiple service resources, resource utilization is low, and response is slow.

[0013] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative work belong to the scope of protection of the application.

[0014] It should be noted that the terms "first", "second" and the like in the specification and the above drawings of the application are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices.

[0015] Embodiment one, as shown in the application, a dynamic optimization allocation method based on power network service resources is provided, the method comprises: Figure 1

[0016] P10: unifying the multi-service resources of the power network into energy flow, and abstracting as a resource space, wherein the multi-service resources at least include power, computing power, storage and communication.

[0017] Further, the embodiment P10 of the application further comprises:

[0018] ​P11: determine the first service resource, call the static resource distribution and dynamic flow transfer mode of the first service resource in the power network as the first resource architecture; P12: introduce the energy flow framework, transform the first resource architecture, and determine the first resource space; P13: traverse the multiple service resources for energy flow unification until the Nth resource space is determined, fuse the first resource space to the Nth resource space, and determine the resource space, wherein N is the number of service resource categories.

[0019] It should be understood that the multiple service resources of the power network are unified by energy flow, and are abstracted into a resource space. The multiple service resources at least include power, computing power, storage, and communication. These resources have different physical properties and functional characteristics in the power network, so they need to be integrated into a resource space through a unified quantification framework for subsequent resource allocation and management.

[0020] First, the first service resource, such as the power resource, is determined, and the static resource distribution and dynamic flow transfer mode of the resource in the power network are called as the first resource architecture. The static resource distribution of the power resource includes power stations, substations, transmission lines, etc., while the dynamic flow transfer mode involves the transmission, distribution and consumption of power. By recording these information in detail, a complete resource architecture of the first service resource in the power network can be constructed, providing a basis for subsequent energy flow unification.

[0021] Next, the energy flow framework is introduced to transform the first resource architecture. The energy flow framework is a standardized quantification method that quantifies different types of resources according to a unified standard for comparison and management in the same dimension. Specifically, for power resources, their power, energy transmission efficiency, etc. can be measured and converted into energy flow units (such as kilowatt hours / second), so that the physical characteristics of power resources are converted into unified quantification indicators. In this way, the first resource architecture is transformed into the first resource space, which is a resource representation characterized by energy flow.

[0022] Subsequently, the other service resources (such as computing power, storage, and communication) in the power network are traversed and unified by energy flow one by one. Taking the computing power resource as an example, its distribution and flow transfer mode in the power network are different from those of the power resource, but it can also be quantified by the energy flow framework. The computing power resource can be converted into an energy flow indicator by calculating the computing power (such as the number of operations per second), thereby constructing the second resource space. Similarly, the storage resource can be quantified by storage capacity and read / write speed, and the communication resource can be quantified by bandwidth and transmission delay, thereby constructing the third resource space and the fourth resource space, respectively.

[0023] After completing the energy flow unification of all service resources, the first resource space to the Nth resource space is fused to determine the complete resource space. N is the number of service resource categories. This fusion process needs to consider the mutual relationship and synergistic effect between different resources to ensure the integrity and consistency of the resource space. For example, the supply of power resources may affect the operating efficiency of computing resources, and the bandwidth of communication resources may limit the data transmission speed of storage resources. By considering these factors comprehensively, a comprehensive and coordinated resource space can be constructed, providing a solid foundation for subsequent resource allocation and management.

[0024] P20: upload the target service task to the resource management platform, determine the allocation mode set by performing semantic calculation coding, wherein each allocation mode is composed of information flow-value flow.

[0025] Further, the embodiment of the application step P20 further comprises:

[0026] P21: identify the target service task, determine the resource allocation orientation by performing semantic parsing, wherein the resource allocation orientation is flexible demand; P22: according to the resource allocation orientation, perform demand reconstruction under multi-mode derivation to determine the allocation mode set, wherein the information flow is defined on the service side, and the value flow is defined on the allocation side. Wherein, the service side includes a first service side based on the management side and a second service side based on the user side, wherein the service side is extensible.

[0027] Optionally, the target service task is uploaded to the resource management platform, and the allocation mode set is determined by semantic calculation coding. This process not only involves the identification and analysis of the task, but also includes the determination of the resource allocation orientation and the construction of the allocation mode.

[0028] First, upload the target service task to the resource management platform. This platform is a centralized management system for handling and optimizing resource allocation tasks in the power network. The target service task can be any service request in the power network, such as power supply, data processing, storage demand, or communication tasks, etc. These tasks usually have complex characteristics, including the amount of resource required, priority, execution time window, and dependency relationship with other tasks, etc.

[0029] After uploading the task, the target service task is first identified and the resource allocation orientation is determined through semantic profiling. Semantic profiling is a technical means to deeply understand the semantic content of the service task, so as to extract the key features and requirements of the task. For example, through natural language processing technology or specific semantic analysis algorithms, the type, quantity, priority and other information of the resource demand implied in the task can be identified. In this process, the system determines that the resource allocation orientation is flexible demand, that is, the demand for resources of the task is not fixed, but can be dynamically adjusted according to the availability of resources and the priority of the task. The identification of this flexible demand provides flexibility and adaptability for subsequent resource allocation.

[0030] Next, demand restructuring under multi-mode derivation is performed according to the resource allocation orientation to determine the allocation mode set. The allocation mode set is the core of the resource allocation strategy, which defines how to allocate resources in different situations to meet the demand of the service task. Each allocation mode is composed of an information flow and a value flow. The information flow defines the transmission path and method of data in the task execution process, while the value flow defines the priority and benefit target of resource allocation. In this application, the service end includes a first service end based on the management side and a second service end based on the user side, and the service end is extensible. This means that the system can dynamically adjust the definition of information flow and value flow according to different service end requirements.

[0031] Specifically, the first service end (management side) may focus more on the efficient use of resources and the overall performance of the system, so its information flow definition may focus on optimizing the path and efficiency of resource allocation, and the value flow definition may focus on the stability and reliability of the system. The second service end (user side) may focus more on the rapid response of the task and the user experience, so its information flow definition may focus on the immediacy and interactivity of the task, and the value flow definition may focus on user satisfaction and the timeliness of task completion.

[0032] Through multi-mode derivation, the demand of the task can be restructured according to different service end requirements and resource allocation orientations, thereby generating multiple allocation modes. These allocation modes not only consider the flexible demand of the task, but also consider the specific requirements of different service ends. Collecting these allocation modes into a complete allocation mode set can provide a basis for subsequent resource allocation decisions.

[0033] P30: The resource allocation plug-in embedded in the driving platform identifies the allocation mode set, slices and reconstructs the resource space under each allocation mode, takes the value flow as the decision orientation, executes multi-thread allocation decisions under mode parallelism, performs allocation result fitting, determines the resource allocation strategy, and manages the service resources of the power network.

[0034] Further, the embodiment of the present application further includes P30a to construct a resource allocation plug-in, P30a further includes:

[0035] P31a: embedding the resource space, reconstructing as a first node with dynamic slicing of the resource space, reconstructing the value flow constraint allocation of the resource space as a second node, deploying the resource allocation plug-in; P32a: calling the power grid resource service record, performing data integration based on the cascade of the first node and the second node, as sample data, supervising and training the resource allocation plug-in to convergence.

[0036] Specifically, the resource allocation plug-in embedded in the driving platform is used to realize the dynamic optimization allocation of power network service resources. The plug-in identifies the allocation mode set, performs dynamic slicing reconstruction on the resource space, and executes multi-thread allocation decision-making with value flow as the decision-making guide. Specifically, the resource allocation plug-in divides the resource space into multiple subspaces according to the information flow definition in the allocation mode set, and each subspace corresponds to a specific allocation mode. This slicing reconstruction process allows the resource allocation plug-in to flexibly adjust the allocation mode of resources according to different task requirements and priorities.

[0037] Specifically, to build the resource allocation plug-in, first, the resource space is embedded, and the dynamic slicing reconstruction of the resource space is taken as the first node to ensure the flexibility and adaptability of resource allocation. At the same time, the value flow constraint allocation of the reconstructed resource space is taken as the second node to ensure the maximization of resource allocation benefits. Specifically, the first node is responsible for dynamically slicing the resource space according to the information flow definition in the allocation mode set, generating multiple subspaces. Each subspace corresponds to a specific allocation mode and can meet the needs of different tasks. The second node performs resource allocation constraints on each subspace according to the value flow definition, ensuring that the priority and benefit target of resource allocation are achieved. By deploying the resource allocation plug-in, the functions of the first node and the second node are integrated into a unified framework, providing strong technical support for resource allocation.

[0038] Then, calling the power grid resource service record is the basis for training the resource allocation plug-in. The power grid resource service record contains historical resource allocation data, which reflects the past resource allocation patterns, task requirements, and allocation results. By performing data integration based on the cascade of the first node and the second node, sample data can be generated. These sample data are used for supervised training of the resource allocation plug-in, so that it can learn the optimal resource allocation strategy. The supervised training process compares the allocation results of the plug-in with the actual optimal allocation results, and continuously adjusts the parameters of the plug-in until the training converges, obtaining the resource allocation plug-in. Training convergence means that the plug-in has been able to accurately generate a resource allocation strategy that meets the value flow constraints according to the allocation mode set and the dynamic slicing reconstruction of the resource space.

[0039] Through the above steps, dynamic optimization allocation of power network service resources is realized. Dynamic slicing reconstruction of resource space and value flow constraint allocation provide flexibility and benefit orientation for resource allocation, not only improving the efficiency and adaptability of resource allocation, but also providing important technical support for efficient management of power network.

[0040] Further, the resource space is sliced and reconstructed according to the information flow under each allocation mode, and the multi-thread allocation decision under parallel execution mode is executed with value flow as the decision orientation. The step P30 of the embodiment of the application further includes:

[0041] 31: identifying the set of allocation modes, dividing the processing domain of the resource allocation plug-in according to the number of allocation modes, and determining M resource allocation areas; P32: importing the set of allocation modes into the M resource allocation areas, and performing resource space reconstruction based on the first node in parallel with resource allocation based on the second node to determine M resource allocation results; P33: determining the resource allocation strategy according to the M resource allocation results.

[0042] It should be understood that the process of slicing and reconstructing the resource space according to the information flow under each allocation mode, and the multi-thread allocation decision under parallel execution mode with value flow as the decision orientation can be further refined to improve the efficiency of resource allocation and realize dynamic optimization allocation of power network service resources.

[0043] First, the set of allocation modes is identified, which contains all possible allocation modes, each mode consisting of information flow and value flow. According to the number of allocation modes, the processing domain of the resource allocation plug-in is divided, and M resource allocation areas are determined. For example, if the set of allocation modes contains M different allocation modes, the system divides the resource allocation plug-in into M processing domains, each responsible for processing one allocation mode. This division ensures that each processing domain can independently perform resource allocation tasks, improving the parallel processing capability and efficiency of the system.

[0044] Next, the set of allocation modes is imported into the M resource allocation areas. Each resource allocation area independently performs resource space reconstruction and resource allocation according to its corresponding allocation mode. Specifically, each resource allocation area first reconstructs the resource space based on the first node (dynamic slicing reconstruction of resource space) to generate a subspace suitable for the allocation mode. Then, based on the second node (value flow constraint allocation), resource allocation is performed to ensure that the allocation result meets the priority and benefit target of the value flow. This process is executed in parallel, i.e., M resource allocation areas simultaneously perform resource space reconstruction and resource allocation, greatly improving the efficiency of resource allocation. Finally, each resource allocation area generates a resource allocation result, resulting in M resource allocation results.

[0045] Finally, based on the M resource allocation results, a comprehensive analysis and comparison of the pros and cons of these results is conducted to determine the optimal resource allocation strategy. This process may involve evaluating the allocation results, such as comparing resource utilization, task completion time, user satisfaction, and other indicators of different allocation results. By considering these indicators comprehensively, the system selects one or more optimal allocation results and integrates them into the final resource allocation strategy. This strategy will be applied to the resource management of the power network, ensuring efficient allocation and utilization of resources.

[0046] Throughout the process, the allocation mode set is a set of resource allocation modes generated by the system based on the semantic analysis and coding of the target service task, and each allocation mode is composed of information flow and value flow. The information flow defines the data transmission path and method in the task execution process, and the value flow defines the priority and benefit target of resource allocation. The processing domain division divides the resource allocation plug-in into multiple independent processing units, each of which is responsible for processing one allocation mode, thereby improving the parallel processing capability and efficiency of the system. The resource space reconstruction is dynamically sliced according to the information flow of the allocation mode to generate a subspace suitable for the allocation mode, ensuring the flexibility and adaptability of resource allocation. The value flow constraint allocation is based on the value flow definition of the allocation mode to allocate resources, ensuring that the allocation result meets the goal of maximizing benefits. The multi-threaded allocation decision further improves the efficiency of resource allocation by processing multiple resource allocation tasks in parallel.

[0047] Further, the step P33 of the embodiment of the present application further includes:

[0048] P33-1: mapping the M resource allocation results to allocate consistency, ranking the allocation results, and determining a multi-level classification sequence; P33-2: for the multi-level classification sequence, performing freedom identification and integration to determine the resource allocation strategy.

[0049] Specifically, the process of determining the resource allocation strategy can be further refined. After completing the parallel processing of the M resource allocation regions, the system will obtain M resource allocation results. First, these allocation results are mapped to allocate consistency as the basis for ranking the allocation results. Allocation consistency refers to the degree of similarity or consistency between different allocation results, which reflects the matching degree of each allocation result in the resource allocation strategy. By calculating the allocation consistency, the system can divide the M resource allocation results into multiple levels, forming a multi-level classification sequence. For example, allocation results with higher consistency are classified as high-level, while allocation results with lower consistency are classified as low-level. This classification method helps the system quickly identify which allocation results are closer to the optimal solution, thereby providing a reference for subsequent strategy formulation.

[0050] Subsequently, the system performs degree of freedom identification and integration for the multi-level classification sequence. Degree of freedom identification refers to the evaluation of flexibility and adjustability of each allocation result. Specifically, the system analyzes the variability and adaptability of each allocation result in resource allocation. For example, some allocation results may have high flexibility in resource allocation and can be adjusted according to actual conditions; while other allocation results may be fixed and difficult to modify. Through degree of freedom identification, the system can determine the adjustment space and optimization potential of each allocation result.

[0051] After completing the degree of freedom identification, the system integrates the allocation results in the multi-level classification sequence. The integration process is a comprehensive analysis of allocation results at different levels, combined with the results of degree of freedom identification, to determine the final resource allocation strategy. For example, for high-level allocation results, the system may directly adopt their allocation scheme; while for low-level allocation results, the system may make appropriate adjustments according to the degree of freedom identification to improve their consistency and adaptability. Ultimately, through this hierarchical, identification and integration process, the system can determine a resource allocation strategy that meets the value flow constraints and has high consistency and flexibility, thereby achieving efficient management of power network service resources.

[0052] Throughout the process, allocation consistency is an important indicator of the similarity of different allocation results, which provides a basis for the system's classification; degree of freedom identification provides flexibility for the system's strategy adjustment, enabling the system to optimize allocation results according to actual conditions. Through this refined processing method, the accuracy and adaptability of resource allocation can be effectively improved.

[0053] Further, the step P33-1 of the embodiment of the present application further includes:

[0054] P33-11: Take the allocation part consistent between M items in M resource allocation results as a first-level classification sequence; P33-12: Take the allocation part consistent between M-1 items in M resource allocation results as a second-level classification sequence; P33-13: Iteratively determine M-level classification sequences, integrate the first-level classification sequence, the second-level classification sequence, and so on until the M-level classification sequence, to determine the multi-level classification sequence, wherein the M-level classification sequence is the allocation part without overlap between M items.

[0055] In a possible embodiment of the present application, the process of allocation result classification can be further refined to implement the hierarchical processing of M resource allocation results and form a multi-level classification sequence.

[0056] After completing the parallel processing of M resource allocation zones, the system obtains M resource allocation results. To hierarchically process these allocation results, first, take the allocation parts consistent among the M items in the M resource allocation results as the first-level classification sequence. This means that the system will identify the parts that are completely consistent among all M allocation results, which are considered the highest priority allocation scheme because they are consistently processed under all allocation modes.

[0057] Next, take the allocation parts consistent among M-1 items in the M resource allocation results as the second-level classification sequence. This step identifies the parts that are consistent among M-1 allocation results, which, although not completely consistent under all allocation modes, are consistent in most cases and therefore have a higher priority.

[0058] Next, continue to iterate sequentially to determine M-level classification sequences. Specifically, the system will gradually reduce the number of consistent items, starting from allocation parts consistent among M-2 items, and gradually reducing to allocation parts without overlap among M items. Finally, the system will form a complete multi-level classification sequence, with M-level classification sequences being allocation parts without overlap among M items. These non-overlapping parts are considered the lowest priority allocation scheme because they have significant differences under different allocation modes.

[0059] By integrating the first-level classification sequence, the second-level classification sequence, and up to the M-level classification sequence, the system ultimately determines the multi-level classification sequence. This sequence not only reflects the degree of consistency between different allocation results, but also provides a clear hierarchical basis for subsequent resource allocation strategy formulation. For example, the system can prioritize allocation schemes in the first-level classification sequence because they are consistently recognized under all allocation modes; for lower-level classification sequences, the system can adjust and optimize according to actual conditions.

[0060] Throughout the process, through step-by-step iteration, starting from the most consistent allocation parts, the consistency requirements are gradually reduced, and a comprehensive multi-level classification sequence is ultimately formed. This method not only ensures the flexibility and adaptability of resource allocation strategies, but also improves the accuracy and efficiency of resource allocation. Through this fine-grained hierarchical processing, complex and variable resource allocation demands in power networks can be effectively addressed, improving resource utilization and allocation accuracy in power network resource management.

[0061] Further, the step P30 of the embodiments of the present application further includes:

[0062] P34: The classification sequence level is positively correlated with the degree of freedom, and the power grid resource management is non-uniformly regulated according to the degree of freedom identifier and the resource allocation strategy execution scale.

[0063] Specifically, after determining the multi-level classification sequence, the system performs non-uniform regulation scale power grid resource management according to the positive correlation between the classification sequence level and the degree of freedom. Specifically, the higher the level of the classification sequence, the stronger the adaptability of the allocation result under different modes, and the higher the degree of freedom; on the contrary, the lower the level, the lower the degree of freedom. This positive correlation provides an important guide for resource management.

[0064] First, the same task under different modes is analyzed in parallel to determine multiple allocation results. These results reflect the resource allocation scheme for the same task under different allocation modes. For example, for a specific service task, the system may generate different allocation results under multiple modes (such as high priority mode, low delay mode, high reliability mode, etc.).

[0065] Next, the consistency of these allocation results is analyzed. If there is a consistent allocation part in multiple results, it indicates that this part of resource allocation is suitable for multiple modes. For these consistent allocation parts, the minimum degree of freedom constraint is adopted, that is, strictly according to the allocation result. This is because these parts have been consistently recognized under multiple modes, have high stability and reliability, and can therefore be used as the basis for resource allocation.

[0066] However, if there are inconsistent allocation parts in multiple results, it indicates that these parts have weak adaptability and high individuality. In this case, the actual uncontrollable factors and stability are limited, and the corresponding adaptive regulation demand is also higher. Therefore, during the implementation of the allocation, dynamic constraints need to be made on these parts. Specifically, the system will dynamically adjust the resource allocation of these parts according to the degree of freedom identifier and resource allocation strategy to ensure the flexibility and adaptability of resource allocation. For example, for an allocation part that is consistent in multiple modes, the system may directly adopt the allocation scheme; for an allocation part that is consistent in only a few modes, the system may adjust it according to the actual situation, such as increasing the redundancy of resources or adjusting the priority of resources, to cope with uncontrollable factors and improve stability.

[0067] Through the positive correlation between the classification sequence level and the degree of freedom, non-uniform regulation scale power grid resource management can be achieved. This method not only ensures the flexibility and adaptability of resource allocation, but also improves the accuracy and efficiency of resource allocation. At the same time, through dynamic constraints and adjustments, complex and variable resource allocation demands in the power network can be effectively coped with, and the accuracy and response speed of dynamic resource management of the power network can be improved.

[0068] In summary, the embodiments of the present application have at least the following technical effects:

[0069] The application solves the problem of resource heterogeneity through the energy flow unification of various heterogeneous resources, improves management flexibility, quickly responds to task demand changes through semantic calculation coding and multi-thread allocation decision, improves resource allocation real-time performance and adaptability, realizes resource allocation benefit maximization based on information flow slicing reconstruction and value flow oriented decision, improves resource utilization efficiency, adopts dynamic constraints and adaptive control, enhances system stability and reliability, reduces the influence of uncontrollable factors, realizes resource allocation intelligence and automation, and provides technical support for efficient management of power network.

[0070] The technical effects of efficient management of multi-service resources, improved resource utilization and response speed are achieved through energy flow unification and dynamic optimization allocation.

[0071] Embodiment two, based on the same inventive concept as the dynamic optimization allocation method of power network service resources in the foregoing embodiments, as shown in the following table, the application provides a dynamic optimization allocation system based on power network service resources, and the system and method embodiments in the application embodiment are based on the same inventive concept. Wherein, the system comprises: Figure 2

[0072] The energy flow unification module 11 is used for energy flow unification of multi-service resources of the power network, and is abstracted as a resource space, wherein the multi-service resources at least include power, computing power, storage and communication.

[0073] The allocation mode determination module 12 is used for uploading target service tasks to the resource management platform, determining the allocation mode set by performing semantic calculation coding, wherein each allocation mode is composed of information flow-value flow.

[0074] The multi-thread allocation decision module 13 is used for driving the resource allocation plug-in embedded in the platform, identifying the allocation mode set, slicing the resource space under each allocation mode, taking the value flow as the decision orientation, executing the multi-thread allocation decision under the mode parallel, fitting the allocation result, determining the resource allocation strategy, and managing the service resources of the power network.

[0075] Further, the energy flow unification module 11 is further used for performing the following steps:

[0076] ​Determine the first service resource, call the static resource distribution and dynamic flow transfer mode of the first service resource in the power network as the first resource architecture; introduce the energy flow framework, transform the first resource architecture, and determine the first resource space; traverse the multiple service resources to unify the energy flow, until the Nth resource space is determined, fuse the first resource space to the Nth resource space, and determine the resource space, wherein N is the number of service resource categories.

[0077] Further, the allocation mode determination module 12 is further used to execute the following steps:

[0078] Identify the target service task, determine the resource allocation orientation by performing semantic parsing, wherein the resource allocation orientation is flexible demand; according to the resource allocation orientation, perform demand reconstruction under multiple modes, and determine the allocation mode set, wherein the service end defines information flow and the allocation side defines value flow. The service end includes a first service end based on the management side and a second service end based on the user side, wherein the service end is extensible.

[0079] Further, the multi-threaded allocation decision module 13 is further used to execute the following steps:

[0080] Embed the resource space, reconstruct the first node with the dynamic slice of the resource space, reconstruct the value flow constraint allocation of the resource space as the second node, and deploy the resource allocation plug-in; call the power grid resource service record, perform data integration based on the cascade of the first node and the second node as sample data, and supervise and train the resource allocation plug-in to convergence.

[0081] Further, the multi-threaded allocation decision module 13 is further used to execute the following steps:

[0082] Identify the allocation mode set, divide the processing domain of the resource allocation plug-in according to the number of allocation modes, and determine M resource allocation areas; import the allocation mode set into the M resource allocation areas, and perform resource space reconstruction based on the first node and resource allocation based on the second node in parallel to determine M resource allocation results; according to the M resource allocation results, determine the resource allocation strategy.

[0083] Further, the multi-threaded allocation decision module 13 is further used to execute the following steps:

[0084] Map the M resource allocation results to the allocation consistency degree, perform allocation result grading, and determine a multi-level classification sequence; for the multi-level classification sequence, perform freedom degree identification and integration to determine the resource allocation strategy.

[0085] Further, the multi-threaded allocation decision module 13 is further used to execute the following steps:

[0086] Taking the consistent allocation parts between M-1 items in the M resource allocation results as the second-level classification sequence; iteratively determining the M-level classification sequence, and integrating the first-level classification sequence, the second-level classification sequence and the M-level classification sequence to determine the multi-level classification sequence, wherein the M-level classification sequence is the allocation part without overlapping between M items.

[0087] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. Moreover, the above-mentioned embodiments of the present application are described. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are possible or can be advantageous.

[0088] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0089] The specification and drawings of the present application are only exemplary descriptions of the present application, and are considered to cover any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalent technology, the present application intends to include these modifications and changes.

Claims

1. A dynamic optimization allocation method for power network service resources, characterized in that, The method includes: The power grid's multi-service resources are unified in terms of energy flow and abstracted into a resource space. These multi-service resources include at least power-related, computing-related, storage-related, and communication-related resources. Upload the target service task to the resource management platform, and determine the allocation pattern set by performing semantic decoding and encoding. Each allocation pattern consists of information flow and value flow. The resource allocation plugin embedded in the driving platform identifies the allocation mode set, reconstructs the resource space by slicing it with the information flow of each allocation mode, uses the value flow as the decision guide, executes multi-threaded allocation decisions under parallel execution modes, fits the allocation results, determines the resource allocation strategy, and manages the service resources of the power network.

2. The method as described in claim 1, characterized in that, The energy flow of multiple service resources in the power network is unified and abstracted into a resource space, including: Determine the first service resource, retrieve the static resource distribution and dynamic flow mode of the first service resource in the power network, and use it as the first resource architecture; An energy flow framework is introduced to transform the first resource architecture and determine the first resource space; The multiple service resources are traversed for energy flow unification until the Nth resource space is determined. The first resource space up to the Nth resource space are merged to determine the resource space, where N is the number of service resource categories.

3. The method as described in claim 1, characterized in that, By performing semantic decoding and encoding, the allocation pattern set is determined, including: Identify the target service task, and determine the resource allocation orientation by performing semantic analysis, wherein the resource allocation orientation is elastic demand; Based on the resource allocation guideline, demand restructuring under multi-mode derivation is performed to determine the allocation mode set, wherein the information flow is defined by the server and the value flow is defined by the allocation focus.

4. The method as described in claim 3, characterized in that, The server includes a first server based on the management side and a second server based on the user side, wherein the server is scalable.

5. The method as described in claim 1, characterized in that, Before the resource allocation plugin is embedded in the driver platform, the construction of the resource allocation plugin includes: The resource space is embedded, with the dynamic slicing and reconstruction of the resource space as the first node and the value stream constraint allocation of the reconstructed resource space as the second node, and the resource allocation plugin is deployed. Retrieve power grid resource service records, perform data integration based on the cascading of the first and second nodes, and use the data as sample data to conduct supervised training of the resource allocation plugin until convergence.

6. The method as described in claim 5, characterized in that, The resource space is sliced ​​and reconstructed based on the information flow under each allocation mode. Value stream-driven decision-making guides multi-threaded allocation decisions under parallel execution modes, including: Identify the allocation pattern set, and divide the resource allocation plugin into processing domains according to the number of allocation patterns to determine M resource allocation areas; The allocation pattern set is imported into the M resource allocation areas, and resource space reconstruction based on the first node and resource allocation based on the second node are performed in parallel to determine the M resource allocation results. Based on the M resource allocation results, a resource allocation strategy is determined.

7. The method as described in claim 6, characterized in that, Based on the M resource allocation results, a resource allocation strategy is determined, including: Map the M resource allocation results, classify the allocation results according to the allocation consistency, and determine a multi-level classification sequence; For the multi-level classification sequence, the degrees of freedom are identified and integrated to determine the resource allocation strategy.

8. The method as described in claim 7, characterized in that, Take the consistent allocation portions among the M items in the M resource allocation results as the first-level classification sequence; Take the consistent allocation portion among the M-1 items in the M resource allocation results as the secondary classification sequence; The process is iterated sequentially to determine the M-level classification sequence. The first-level classification sequence, the second-level classification sequence, and so on up to the M-level classification sequence are integrated to determine the multi-level classification sequence. The M-level classification sequence is the non-overlapping allocation part among the M items.

9. The method as described in claim 1, characterized in that, The classification sequence level is positively correlated with the degree of freedom, and power grid resource management with unbalanced control scale is performed according to the degree of freedom identifier and the resource allocation strategy.

10. A dynamic optimization allocation system based on power network service resources, characterized in that, The system is used to perform the method according to any one of claims 1 to 9, the system comprising: The power flow unification module is used to unify the power flow of multiple service resources of the power network, which is abstracted into a resource space. The multiple service resources include at least power, computing power, storage and communication resources. The allocation mode determination module is used to upload the target service task to the resource management platform and determine the allocation mode set by performing semantic decoding and encoding. Each allocation mode consists of an information flow and a value flow. A multi-threaded allocation decision module is used to drive the resource allocation plugin embedded in the platform, identify the allocation mode set, reconstruct the resource space by slicing with the information flow of each allocation mode, take the value flow as the decision guide, execute multi-threaded allocation decisions under parallel execution modes, fit the allocation results, determine the resource allocation strategy, and perform service resource management for the power network.

Images